About Local Climate Trends
Understanding our mission and approach to climate data analysis
Author
Local Climate Trends was created by Ethan Coffel, a climate scientist at Syracuse University. This platform is designed to make high-quality climate data easily accessible.
Data sources
Our analysis uses NOAA's nClimGrid-Daily dataset, which provides high-resolution, gridded climate data for the contiguous United States. This dataset is based on daily temperature and precipitation records from thousands of weather stations and spans from 1951 to the present, designed to support climate monitoring, research, and impact assessments at local to regional scales. NOAA averages it over every county, and the current year is updated daily from NOAA's preliminary numbers, about three days behind.
Some measures need more than temperature and precipitation:
- Humidity (muggy days, nice days, the summer heat index) and wind come from gridMET (University of California, Merced), from 1979. gridMET's humidity is adjusted to match dew points measured at about 400 airport weather stations (NOAA ISD and the Iowa Environmental Mesonet).
- Snowfall is measured snowfall at NOAA weather stations (GHCN-Daily), from 1980. Only stations with nearly complete records are used, from the county itself or up to three within 50 km (31 miles). Counties without such a station, or with very little snow, are left blank.
- Fire weather uses the Canadian Fire Weather Index, worked out each day from nClimGrid temperature and precipitation and gridMET humidity and wind.
- County outlines, places and ZIP codes come from the US Census Bureau.
Methodology
We employ statistical regression analysis to identify long-term climate trends. The change shown for each county is the least-squares linear trend over 1979–2025, expressed as the fitted value in 2025 minus the fitted value in 1979. Counties are hatched where the trend is not statistically significant at the 5% level (two-sided t-test on the slope, with the effective sample size adjusted for lag-1 autocorrelation following Santer et al. 2000).
The year-by-year view shows each year against the 1951–1980 average. Each county page also shows how the seasons, months and records have changed, and compares day-by-day climate in 1996–2025 with 1951–1980.
A county's climate twin is the county whose climate in 1951–1980 most closely matches its own climate in 1996–2025, measured by the standardized Euclidean distance (Williams, Jackson and Kutzbach 2007, PNAS) over seasonal mean daily maximum and minimum temperature and total precipitation, with each difference divided by the county's 1951–1980 interannual standard deviation.
Features
Explore a map of every county in the Lower 48 with more than a dozen measures, from temperature and hot days to freezing nights, snowfall, heavy rain, humidity and fire weather. Search for any city, county or ZIP code, open a county for detailed charts and records, compare two counties side by side, and download any county's numbers as a spreadsheet (CSV).
Free to use
All maps, charts and data on Local Climate Trends are free to download and use for any purpose. No attribution is required, though it is always appreciated.